Model: SoLU Model: 10 Layers, 5120 Neurons per Layer

Dataset: 80% C4 (Web Text) and 20% Python Code

Neuron 704 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-10l')

Text #0

Max Range: 1.1138. Min Range: -1.1138

Max Act: 1.1138. Min Act: -0.0001

Data Index: 258006 (C4 (Web Text))

Max Activating Token Index: 599

Text #1

Max Range: 1.1138. Min Range: -1.1138

Max Act: 1.0484. Min Act: -0.0001

Data Index: 1488279 (Python Code)

Max Activating Token Index: 881

Text #2

Max Range: 1.1138. Min Range: -1.1138

Max Act: 1.0649. Min Act: -0.0001

Data Index: 863270 (C4 (Web Text))

Max Activating Token Index: 398

Text #3

Max Range: 1.1138. Min Range: -1.1138

Max Act: 0.9997. Min Act: -0.0001

Data Index: 795013 (C4 (Web Text))

Max Activating Token Index: 476

Text #4

Max Range: 1.1138. Min Range: -1.1138

Max Act: 1.0488. Min Act: -0.0001

Data Index: 1034265 (C4 (Web Text))

Max Activating Token Index: 982

Text #5

Max Range: 1.1138. Min Range: -1.1138

Max Act: 0.8989. Min Act: -0.0001

Data Index: 580333 (C4 (Web Text))

Max Activating Token Index: 245

Text #6

Max Range: 1.1138. Min Range: -1.1138

Max Act: 0.9041. Min Act: -0.0001

Data Index: 1106558 (C4 (Web Text))

Max Activating Token Index: 940

Text #7

Max Range: 1.1138. Min Range: -1.1138

Max Act: 0.8976. Min Act: -0.0001

Data Index: 775968 (C4 (Web Text))

Max Activating Token Index: 920

Text #8

Max Range: 1.1138. Min Range: -1.1138

Max Act: 0.8042. Min Act: -0.0001

Data Index: 435025 (C4 (Web Text))

Max Activating Token Index: 547

Text #9

Max Range: 1.1138. Min Range: -1.1138

Max Act: 0.8495. Min Act: -0.0001

Data Index: 16994 (C4 (Web Text))

Max Activating Token Index: 737

Text #10

Max Range: 1.1138. Min Range: -1.1138

Max Act: 0.8206. Min Act: -0.0001

Data Index: 250286 (C4 (Web Text))

Max Activating Token Index: 335

Text #11

Max Range: 1.1138. Min Range: -1.1138

Max Act: 0.8490. Min Act: -0.0001

Data Index: 161468 (C4 (Web Text))

Max Activating Token Index: 311

Text #12

Max Range: 1.1138. Min Range: -1.1138

Max Act: 0.8192. Min Act: -0.0001

Data Index: 1013089 (C4 (Web Text))

Max Activating Token Index: 1014

Text #13

Max Range: 1.1138. Min Range: -1.1138

Max Act: 0.8377. Min Act: -0.0001

Data Index: 1006639 (C4 (Web Text))

Max Activating Token Index: 623

Text #14

Max Range: 1.1138. Min Range: -1.1138

Max Act: 0.8222. Min Act: -0.0001

Data Index: 215302 (C4 (Web Text))

Max Activating Token Index: 186

Text #15

Max Range: 1.1138. Min Range: -1.1138

Max Act: 0.8487. Min Act: -0.0001

Data Index: 346994 (C4 (Web Text))

Max Activating Token Index: 187

Text #16

Max Range: 1.1138. Min Range: -1.1138

Max Act: 0.8084. Min Act: -0.0001

Data Index: 371931 (C4 (Web Text))

Max Activating Token Index: 535

Text #17

Max Range: 1.1138. Min Range: -1.1138

Max Act: 0.8015. Min Act: -0.0001

Data Index: 1300800 (C4 (Web Text))

Max Activating Token Index: 947

Text #18

Max Range: 1.1138. Min Range: -1.1138

Max Act: 0.8250. Min Act: -0.0001

Data Index: 487459 (C4 (Web Text))

Max Activating Token Index: 418

Text #19

Max Range: 1.1138. Min Range: -1.1138

Max Act: 0.8205. Min Act: -0.0001

Data Index: 1457453 (Python Code)

Max Activating Token Index: 400